The cleanlab community has grown over the years. Today, we are excited to release cleanlab 1.0 as the standard package for machine learning with noisy labels and finding errors in datasets.

If you're coming from the research side (e.g. the confident learning or label errors paper) -- use this version of cleanlab.

cleanlab 1.0

cleanlab 1.0 supports the most common versions of python (2, 2.7, 3.4, 3.5, 3.6, 3.7, 3.8.) and operating systems (linux, macOS, Windows). It works with any deep learning or machine learning library by working with model outputs, regardless of where they come from. cleanlab also has built-in support now for new research from other scientists (e.g. Co-Teaching) outside of our group at MIT.

More details about new features of cleanlab 1.0 below:

• Added Amazon Reviews NLP to cleanlab/examples
• cleanlab now supports python 2, 2.7, 3.4, 3.5, 3.6, 3.7, 3.8.
• Users have used cleanlab with python version 3.9 (use at your own risk!)
• Added more testing. All tests pass on windows/linux/macOS.
• Update to GNU GPL-3+ License.
• Added documentation: https://cleanlab.readthedocs.io/
• The cleanlab "confident learning" paper is published in the Journal of AI Research: https://jair.org/index.php/jair/article/view/12125
• Added funding, community and contributing guidelines
• Fixed several errors in cleanlab/examples
• cleanlab now supports Windows, macOS, Linux, and unix systems
• Many examples added to the README and docs
• cleanlab now natively supports Co-Teaching for learning with noisy labels (reqs python3, PyTorch 1.4)
• cleanlab built in support with handwritten datasets (besides MNIST)
• cleanlab built in support for CIFAR dataset
• Multiprocessing fixed for windows systems
• Adhered all core modules to PEP-8 styling.
• cleanlab is now installable via conda (besides pip).
• Extensive benchmarking of cleanlab methods published.
• Cleanlab now provides future features planned in cleanlab/version.py
• Added confidentlearning-reproduce as a separate repo to reproduce state-of-the-art results.